SMLMFrameConnection

Overview

SMLMFrameConnection performs frame-connection on localization microscopy data organized in an SMLMData.SMLD2D structure (https://github.com/JuliaSMLM/SMLMData.jl). Specifically, SMLMFrameConnection connects repeated localizations of a single blinking event of an emitter into a single higher precision localization. This is done using the algorithm(s) presented in https://doi.org/10.3389/fbinf.2021.724325. Localizations which were connected will share the same unique integer value for the field SMLMData.SMLD2D.connectID.

Interface

Once an SMLMData.SMLD2D structure is fully populated, the user only needs to run a single high-level method from the package: frameconnect(). All fields of the SMLMData.SMLD2D structure must be populated with either meaningful values (e.g., for fields like x, y, σ_x, σ_y, and framenum, which the algorithm depends on) or by placeholders with a meaningful size (e.g., fields like bg and σ_bg, which may not be available, should be set to something like smld.bg = zeros(Float64, length(smld.framenum)), and σ_bg = fill(Inf64, length(smld.framenum))).

The algorithm can be run on the fully populated smld::SMLMData.SMLD2D with default parameters as

smld_connected, smld_preclustered, smld_combined, params = SMLMFrameConnection.frameconnect(smld)

The output smld_connected is a copy of smld with the field smld.connectID updated to associate connected localizations. smld_preclustered is a copy of smld with the field smld_preclustered.connectID updated to associate localizations that belonged to the same precluster. smld_combined contains the combined higher precision localizations (i.e., smld_combined=SMLMFrameConnection.combinelocalizations(smld_connected)) and is considered the main output of this package. params is a structure containing the user-defined and internally-defined parameters used in the algorithm.

Several user-defined parameters can be changed as optional keyword arguments to frameconnect():

smld_connected, smld_preclustered, smld_combined, params = SMLMFrameConnection.frameconnect(smld;
    nnearestclusters = 2, nsigmadev = 5.0,
    maxframegap = 5, nmaxnn = 2)

See the documentation in SMLMFrameConnection/structdefinitions.jl for a description of these parameters and for guidance to which functions use them.

Citation

David J. Schodt and Keith A. Lidke, "Spatiotemporal Clustering of Repeated Super-Resolution Localizations via Linear Assignment Problem", Frontiers in Bioinformatics, 2021 https://doi.org/10.3389/fbinf.2021.724325